Ghost Rating Formula

[TemplarsBane]

I can't figure out WHAT I've done wrong, but I'm working on a little GR Calculator in Google Sheets just to be able to plug in guesses at some stuff and see what outcome I need to increase GR.

But SOMETHING is clearly not working correctly with the math I've got and I don't know where I've screwed up.

In a Classic, full press game, with 7 players, DSS, a 3 way draw, a player GR of 100 and an average of 150 GR for all players (for a total of 1050), what is the expected GR adjustment?

My math is coming back with -14.28 which has to be wrong. A 100 GR player shouldn't LOSE GR in a 3 way draw with players who average a higher score than them.

So any ideas as to what I'm missing?

Here's the calculator if people can spot my mistake in here as well: https://docs.google.com/spreadsheets/d/ ... sp=sharing

[teccles]

I think there's an error in https://webdiplomacy.net/ghostRatings.php - surely the adjustment should be (actual_score - expected_score)*x, not the other way round.

[TemplarsBane]

I'll mess around with that. I feel like I did that and it didn't work right. But I'm trying it now and it DOES seem to work with that...

[Doug7878]

@TemplarsBane
100 GR player contributes 5.714286 (1/17.5 x 100) to the GR pot.
Other players with total GR of 950 contribute 54.285714 (1/17.5 x 950) to the GR pot.
Total contribution to the pot is therefore 60, which is 1050 total GR / 17.5
Three way draw awards 100 GR player 1/3 of 60 = 20, minus his/her contribution of 5.714286 to the GR pot, for a net increase in GR of 54.285714

[gimix]

Three way draw awards 100 GR player 1/3 of 60 = 20, minus his/her contribution of 5.714286 to the GR pot, for a net increase in GR of 54.285714

I would expect a net increase of 20 - 5.71 = 14.29

[TemplarsBane]

Three way draw awards 100 GR player 1/3 of 60 = 20, minus his/her contribution of 5.714286 to the GR pot, for a net increase in GR of 54.285714

I would expect a net increase of 20 - 5.71 = 14.29

With switching Actual - Estimated Scores, 14.29 is what I get as well.